Comparison of admission chest computed tomography and lung ultrasound performance for diagnosis of COVID-19 pneumonia in populations with different disease prevalence

Abstract

Purpose: Chest computed tomography (CT) is considered a reliable imaging tool for COVID-19 pneumonia diagnosis, while lung ultrasound (LUS) has emerged as a potential alternative to characterize lung involvement. The aim of the study was to compare diagnostic performance of admission chest CT and LUS for the diagnosis of COVID-19.

Methods: We included patients admitted to emergency department between February 21-March 6, 2020 (high prevalence group, HP) and between March 30-April 13, 2020 (moderate prevalence group, MP) undergoing LUS and chest CT within 12 h. Chest CT was considered positive in case of "indeterminate"/"typical" pattern for COVID-19 by RSNA classification system. At LUS, thickened pleural line with ≥ three B-lines at least in one zone of the 12 explored was considered positive. Sensitivity, specificity, PPV, NPV, and AUC were calculated for CT and LUS against real-time reverse transcriptase polymerase chain reaction (RT-PCR) and serology as reference standard.

Results: The study included 486 patients (males 61 %; median age, 70 years): 247 patients in HP (COVID-19 prevalence 94 %) and 239 patients in MP (COVID-19 prevalence 45 %). In HP and MP respectively, sensitivity, specificity, PPV, and NPV were 90-95 %, 43-69 %, 96-72 %, 20-95 % for CT and 94-93 %, 7-31 %, 94-52 %, 7-83 % for LUS. CT demonstrated better performance than LUS in diagnosis of COVID-19, both in HP (AUC 0.75 vs 0.51; P < 0.001) and MP (AUC 0.85 vs 0.62; P < 0.001).

Conclusions: Admission chest CT shows better performance than LUS for COVID-19 diagnosis, at varying disease prevalence. LUS is highly sensitive, but not specific for COVID-19.

Keywords: COVID-19; Computed tomography; Diagnostic ultrasound; Spiral.

Fig. 1

Diagram showing the patient selection process. Abbreviations: COVID-19, coronavirus disease 2019; CT, computed tomography; HP, high prevalence; LUS, lung ultrasound; MP, moderate prevalence; RT-PCR, reverse-transcription polymerase chain reaction; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

Fig. 3

A 67 years old male of the early outbreak group admitted to the emergency department for persistent fever and cough from three days. (a) Lung ultrasound performed with convex probe showed in the right lateral lower zone, three B-lines (arrows) and thickened pleural line (arrowheads) considered positive for COVID-19 interstitial pneumonia. (b) Axial unenhanced CT image at the level of inferior pulmonary veins confluence in left atrium, showed pure ground-glass opacities in the middle lobe and right inferior lobe (arrows) while in the left lower lobe depicted a ground-glass opacity with visible intralobular lines, known as “crazy-paving” appearance (arrowheads). Chest CT findings were considered “typical” for COVID-19 pneumonia [20]. Abbreviations: COVID-19, coronavirus disease 2019; CT, computed tomography.

Fig. 2

The 12 zone investigated with lung ultrasound. In the antero-lateral view (a) each hemithorax was divided by the anterior and posterior axillary line in anterior and lateral zone divided by an horizontal zone passing through nipples, identifying four zones: anterior upper zone, anterior lower zone, lateral upper zone, and lateral lower zone. In the postero-lateral view (b) the spinal line, the posterior axillary line and the horizontal nipples lines depicted two additional zone for each hemithorax: posterior upper zone and posterior lower zone. Abbreviations: ALZ, anterior lower zone; AUZ, anterior upper zone; DL, diaphragmatic line; LLZ, lateral lower zone; LUZ, lateral upper zone; PLZ, posterior lower zone; PUZ, posterior upper zone.

Fig. 4

Comparison of ROC curves for diagnosis of COVID-19 in both high prevalence and moderate prevalence groups. Chest CT showed a significant (P < 0.001) better performance in comparison to lung ultrasound in both periods (AUC in the HP 0.75 vs 0.51; AUC in the MP 0.85 vs 0.62). Abbreviations: AUC, area under the ROC curve; COVID-19, coronavirus disease 2019; CT, computed tomography; HP, high prevalence group; MP, moderate prevalence group; ROC, receiver operating characteristic.

Fig. 5

Kaplan-Meier estimates (with 95 % confidence interval) of ICU/death admission for COVID-19 pneumonia extent assessed by LUS and CT. In HP a LUS score >2 (a) and a visual CT score >30 % (b) were significantly associated with shorter ICU admission or death occurrence (both P < 0.001). In MP, a CT visual score>35 % was significantly associated with shorter ICU admission or death occurrence (P < 0.001), while a LUS score >3 (d) failed to identify patients with better prognosis (P = 0.07). Abbreviations: COVID-19, coronavirus disease 2019; CT, computed tomography; HP, high prevalence group; ICU, intensive care unit; LUS, lung ultrasound; MP, moderate prevalence group.